On 22 Oct 2004 at 9:02, Kelly McDonald wrote:

1. Is anyone using this setup in a production environment -

I am using it for a client. Doing development on one machine and then deploying on another, via subversion checkouts.

have there been any problems with stability or performance?

I spent 2 hours this week trying to figure out why, on one machine it (Ape-1.0) worked fine, and on the other it died with an assertion error. Yet the directory contents were identical, the python version the same, all the configs the same. I eventually figured out how to solve the problem by putting an explicit mapper_name=python_script in many "effected" .xyz..properties files. Though, figuring out which one was causing the problem was darn tough! Otherwise it works fine. I don't have any opinion on performance though.

2. Has anyone figured anything out to populate the cache upon startup?

Isn't that a generic ZODB question, not really specific to Ape.. ?

3. What has been the performance hit, if any?

I imagine the loading process is slightly slower than traditional ZODB, but I have not measured it. In our case, we're using it on a site that's all XML and XSLT generation, so there aren't more than 100 or so objects in Ape storage anyway.

4. have you had any problems using subversion/cvs with the objects on the APE/Filesystem volume, especially when using multiple branches for development?

No problems. -- Brad Clements, bkc@murkworks.com (315)268-1000 http://www.murkworks.com (315)268-9812 Fax http://www.wecanstopspam.org/ AOL-IM: BKClements